Eastern Michigan University
DigitalCommons@EMU
Senior Honors Theses
Honors College
2015
The Relation Between Aggressive Behavior and
Engagement in Violence
Andrea Carolina Aya Mercado
Follow this and additional works at: http://commons.emich.edu/honors
Recommended Citation
Aya Mercado, Andrea Carolina, "The Relation Between Aggressive Behavior and Engagement in Violence" (2015). Senior Honors
Theses. 439.
http://commons.emich.edu/honors/439
This Open Access Senior Honors Thesis is brought to you for free and open access by the Honors College at DigitalCommons@EMU. It has been
accepted for inclusion in Senior Honors Theses by an authorized administrator of DigitalCommons@EMU. For more information, please contact [email protected].
The Relation Between Aggressive Behavior and Engagement in Violence
Abstract
Exposure to community violence is positively correlated with aggression towards other people (Allwood &
Bell, 2008). Researchers in this area have classified aggression into two broad categories, proactive and
reactive aggression. Proactive aggression refers to aggression that is used as an instrument to achieve a goal,
whereas reactive aggression refers to aggression that is used as a response to a real or perceived treat (Chaux,
Arboleda, & Rincon, 2012). Exposure to community violence impacts both reactive and proactive aggression,
but mechanisms explaining this relationship are still unclear. One potential mediator of this relationship is
religiosity. Previous research suggests that individuals turn to religion as a coping mechanism when
encountering stressful situations, but the efficacy of this coping mechanism has generated mixed results (Ano
& Vasconcelles, 2005). Considering that exposure to violence and aggressive behavior lead to stress, it is
plausible that religiosity is a mediator of this relationship. One hundred twenty three currently enrolled
Eastern Michigan University undergraduate students completed self-report measures assessing exposure to
community violence, religiosity and both reactive and proactive aggression. It was hypothesized that higher
levels of exposure to violence are positively correlated with both types of aggression and religious coping
would moderate that relationship. This hypothesis was not supported. However, there was a direct effect
between experiences with 'safety' (the mirror image of violence exposure) and lower levels of proactive
aggression.
Degree Type
Open Access Senior Honors Thesis
Department
Psychology
First Advisor
Dr. Dean Lauterbach
Second Advisor
Dr. Stephen Jefferson
Keywords
proactive aggression, reactive aggression, perceptions of safety, violence, religiosity, students
This open access senior honors thesis is available at DigitalCommons@EMU: http://commons.emich.edu/honors/439
Violence and Aggression 1
Running head: AGGRESSIVE BEHAVIOR AND VIOLENCE
The Relation Between Aggressive Behavior and Engagement in Violence
By
Andrea Carolina Aya Mercado
A Senior Thesis Submitted to the
Eastern Michigan University Honors College
In Partial Fulfillment of the Requirements for Graduation
With Honors in Psychology
Approved at Ypsilanti, Michigan, on this date
Violence and Aggression
2
Table of Contents
List of Tables
3
List of Figures
.4
Abstract.
5
The Relation Between Aggressive Behavior and Engagement in Violence
6
Defining Aggression
Frustration-Aggression
7
Theory
8
Cognitive Neoassociation
10
Social Learning Theory
11
Social Information Processing Model
13
Religiosity and Aggression
15
Reactive and Proactive Aggression
16
M~h~
17
Procedures
17
Measures
18
Data analysis
21
Results
21
Participants
21
Bivariate Relationships Among Predictors, Mediators, and outcomes
26
Mediation Analysis
26
Discussion
33
References
37
Violence and Aggression
3
List of Tables
Table 1. Age of the sample
21
Table 2. Year in College
22
Table 3. Racial/Ethnic Heritage of the sample
22
Table 4. Reported Religion
23
Table 5. Measures included
24
Table 6. Mediation Model 1
26
Table 7. Mediation Model 2
27
Table 8. Mediation Model 3
27
Table 9. Mediation Model 4
28
Table 10. Med iation Model 5
28
Table 11. Mediation Model 6
29
Table 12. Mediation Model 7
29
Table 13. Mediation Model 8
30
Violence and Aggression
4
List of Figures
Figure 1. Mediation Analysis
26
Violence and Aggression
5
Abstract
. •.•.•
.'!~ .
Exposure to community violence is positively correlated with aggression towards other
people (Allwood & Bell, 2008). Researchers in this area have classified aggression into
.",:
~ -,J
....::.,
..••.
,.J,...~, 'l;:- ,'~ ..,., ',"-,
"~'.'
two broad categories, proactive and reactive aggression~Proactive aggression refers to
',J
,
...
.
.
.~I.
r, •.
_
,•••
",":
,.
,.•
.~I. "
' ••
..,••
,
.,' ,~
r.
~..
'~
• ':
aggression that is used as an instrument to achieve a goal, whereas reactive
':;
.. ,/
'~~l""".
'
_'
".{
",.,;
" ';:~
••I
',.
".
.
t:.,,;.
••'-.'.~,.'.~
~~',
',"",:
:.~".
I
aggression refers to aggression that is used as a response to a real or perceived treat
,
.:'
:..
t ..
I',r"'
',:
""'."".'
_'
,_
.•
'.,
.•
~~..
<',.
tilt
•
:"(,,""
,',-
":'"
". -
(Chaux, Arboleda, & Rincon, 2012). Exposure to community violence impacts both
,.
-:...~.'
.
";:
.••••.
,f
-','
," "
"::t..
.
,",,'__'"
-~.'"
•.
'.
~
reactive and proactive aggression, but mechanisms explaining this relationship are still
. ".
\
.•..•
- ,~..,
,
';l '_ t
'_~~~.~.
t
~") ~
~:
C1'_'
~
unclear. One potential mediator of this relationship is religiosity. Previous research
.t.; .:.~'
I
t'
.• ,
,>. .,'::
r:",
k.
.~"-.:
.." ',. '.-/
j
'J",'~
,...~.c.: ,1.
~ r.;;
~:
,t,
'.
suggests that individuals turn to religion as a coping mechanism when encountering
..
,:
_ .;'.
'," ~..>
..
, ..:.~.~P''''
.•....
:.i~......•.
t",
~'t
.':"
.t.~.,......'''It,
stressful situations', but the efficacy ofthis coping mechanism has generated mixed
results '(Ano & Vasconcelles, 2005). Considering that exposure to violence and
aggressive behavior lead to stress, it is plausible that religiosity is a mediator of this
relationship. One hundred twenty three currently enrolled Eastern Michigan University
undergraduate students completed self-report measures assessing exposure to
community violence, religiosity and both reactive and proactive aggression. It waS
hypothesized that higher levels of exposure to violence are positively correlated with
both types of aggression
and religious coping would moderate that relationship. This
hypothesis was not supported.
However, there was a direct effect between experiences
with 'safety' (the mirror image of violence exposure) and lower levels of proactive
aggression.
Violence and Aggression
6
Acknowledgements
I would primarily like to thank Dr. Lauterbach who provided continuous encouragement
and support throughout the project. I would like to thank him for all his hard work and
commitment to the thesis as well as his guidance regarding the research process. I
would also like to thank the doctoral students Stefanie Poehacker and David Phillips for
their continuous support in the design and writing process. I am also grateful to Juan
Velandia who supported and encouraged me to make this thesis idea a reality. Finally, I
would like to express my gratitude to the Psychology Department and Eastern Michigan
University for giving me the opportunity and resources to conduct research.
Violence and Aggression
7
The Relation Between Aggressive Behavior and Engagement in Violence
Defining Aggression
Aggression is defined as the combination of different behaviors characterized by
anger, hostility, impulsivity or irritability but can vary in severity and type (Coccaro,
2003). Anderson and Bushman defined aggression as any behavior toward another
individual that intents to cause harm (Chaux, Arboleda, & Rincon, 2012). The intent to
cause harm could be either a response to a provocation or an instrument to achieve a
goal and the two forms of aggression are categorized as reactive and proactive
aggression.
Reactive aggression refers to an aggressive behavior in response to a real
or perceived provocation whereas proactive aggression is used as an instrument to
achieve specific goals without a provocation.
A previous study conducted by Los Andes University in Colombia used two
samples of children from Bogota with different levels of exposure to violence. As
expected, they found that "Exposure to community violence and to youth gangs were
both significantly correlated to reactive and proactive aggression. However, in the
regression analyses, exposure to community violence significantly predicted reactive
and proactive aggression while exposure to youth gangs predicted only proactive
aggression" (Chaux, Arboleda, & Rincon, 2013 p. 243). Therefore the article suggests
that exposure to community violence is a predictor of both types of aggression.
Andreu et al (2003) compared levels of aggression in two samples of Spanish
and Colombian college students. The study sought to determine the influence of
economic and social factors on engagement in aggressive behavior. Andreu et al
Violence and Aggression
8
found that Spaniards engage in reactive aggression while Colombians engage in
proactive aggression. Hence, Colombians behave aggressively to solve problems,
obtain rewards and avoid punishment whereas Spaniards engage in aggressive
behavior when they have no control over the situation or want to express negative
emotions (Andreu, et aI., 2003). Colombian students showed significantly higher scores
in engaging in proactive aggression in comparison with Spaniard students who engaged
in more reactive aggression (24.92 vs. 3.07; p <0.001). Regarding the economic and
social factors Andreu et al. (2003) argue that as Colombia has higher levels of violence
the aggression tends to be proactive while as Spain has lower levels of violence
therefore the aggression is frequently reactive.
The Andreu, et al. study suggests that high exposure to violence is directly
related to the engagement in proactive aggression.
To date, there is not a unified
theory to explain human aggression (Roach, 2002). The following section attempts to
review the main theories that explain the development and permanence of aggressive
behavior within humans from a psychosocial perspective.
Frustration-Aggression
Theory
The frustration-aggression
theory was proposed by John Dollard, Leonard
Doob, Neal Miller, a.H. Mowrer, and Robert Sears in 1939 (Berkowitz, 1989). The
theory was the first to define aggression as "a reaction to environmental conditions,
namely those conditions provoking frustration" (Roach, 2002, p. 2). But before the
theory can be understood it is necessary to define the concepts involved. Frustration is
defined by Dollard et al. as "an interference with the occurrence of an instigated goal-
Violence and Aggression
9
response at its proper time in the behavior sequence" (as cited in Berkowitz, 1989, p.
60). In this context, aggression was defined as a reaction to real or perceived
provocation in the environment.
Based on the definition of frustration, the frustration-
aggression hypothesis proposes that aggression is a response to frustration and it
always emerges from frustration. In the same way "frustration always leads to some
form of aggression" (Ardila, 2005, p.65). In a somewhat more recent version of the
theory, Berkowitz (1962) posited that even though frustration could provoke multiple
responses including aggression, disappointment and depression "aggression will always
be the product of frustration" (as cited in Roach, 2002, p. 3)
The relationship between frustration and aggression has also been examined
in combat veterans.
Perhaps one of the most infamous incidences involving US
combatants in the Vietnam war was the massacre of Vietnamese civilians, including
women and children, at My Lai on March 16th, 1968. The platoon leader was 1st Lt.
William Calley.
Calley was found guilty of killing 22 civilians but his sentence was
reduced to 10 years. Calley was paroled in 1974. Following the frustration aggression
theory an observation of the massacre of My Lai was done. The participants of the My
Lai massacre responded to questionnaires and the results indicated that the level of
aggression of the participants increased due to the fear and frustration provoked by the
war. Therefore "war is not caused by human aggressiveness, or that people fight in war
because they are aggressive" (Feshbach & Zagrodzka, 1997 p. 178) but rather the
frustration emerges from situations of violence and war, which then leads to aggression.
Violence and Aggression
10
Cognitive Neoassociation
The cognitive neoassociation theory expanded the score of the Dollad et al.
frustration aggression theory. It argues that frustration is one of the many conditions
that can provoke negative affect (Roach, 2002) and "any condition that evokes negative
affect may serve as an antecedent of aggression and/or escape/avoidance
behavior"
(Crevecoeur, 2007 p. 15) that could or could not turn out to be aggressive behavior.
Previous experiments have tried to understand the relationship between
violence and the negative affect caused by physical discomfort. Bell and Baron (1979)
studied the influence of temperature on aggression in a sample of 35 undergraduates.
The participants were assigned a certain task and after completing the task the
participants were given the opportunity to administer an electrical shock to the
evaluator. The study concluded that high temperatures increased the negative affect
due to physical discomfort and the participants became more aggressive (Bell & Baron,
1979). Additionally, other sources of discomfort have been linked to negative affect and
increased aggression such as negative assessments of the self in regards to skills and
intelligence or the explicit aggressiveness of other members of the experiment. These
studies concluded that induced pain increases the likelihood to aggress against an
available target (Berkowitz, 1989)
. Presumably, such aggression results in a decrease
in frustration which in turn strengthens the relationship between negative affect and
behavior.
Violence and Aggression
11
On the other hand, the research methodology applied in the cognitive
neoassociation theory has been criticized due to its inconsistency with the actual
context of human aggression. Critics argue that the laboratory conditions that have
been manipulated in these studies permit aggressive behaviors (e.g., shocking people)
that cannot be generalized to situations outside the laboraory (Crevecoeur, 2007).
Social Learning Theory
The social learning theory emerged with Albert Bandura in the 1960s and
focuses on the process of developing and changing aggressive behaviors. Bandura's
work shifted attention away from frustration, biological factors, and reinforcements as
causes of aggression.
Instead, Bandura looked to explain how aggressive behavior is
learned and focused on "the process of learning through observation or by example ...
either intentionally or accidentally" (Engler, 2014 p. 214).
In a typical study testing Bandura's social learning model, a participant
observed a confederate performing either aggressive behavior (experimental group), or
non-aggressive behavior (control group). Later on the participant is exposed to a
frustrating situation.
Subjects exposed to an aggressive model are compared to
subjects exposed to a non-aggressive model on level of aggression.
The most famous of Bandura's experiment included a Bobo doll in which the
model, an adult, behaved aggressively toward the doll hitting and kicking it. The children
in the experimental group observed the model behaving aggressively while the control
group viewed a non-aggressive model. Afterwards the children in each of the groups
Violence and Aggression
12
were allowed to play with the Bodo doll. As hypothesized, the experimental group that
observed an aggressive model was considerably more aggressive than the control
group. Bandura concluded that observational learning exceeds straightforward imitation
because individuals engage in the observed behavior and interpret the behavior adding
new elements to it. In Bandura's experiments the main distinctions were between
mimicking the model, imitation, and match the style and structure of the behavior,
modeling (Engler, 1999).
Bandura identified three main factors that influence the modeling process. The
first factor consists in the characteristics of the model in relation to the observer, the
greater similarity between the model and the observer, the greater the emulation of the
behavior. Additionally, the aggression modeling process is also influenced by the
complexity and kind of behavior been emulated by the observer (Engler, 1999). When
the model engages in a simple behavior, such as saying please and thank you, the
observer is more likely to model that behavior. On the other hand, if the model engages
in a complex behavior, such as solving a difficult math problem, the observer is less
likely to model that behavior. Arguably, aggressive behavior is fairly simple and
consequently easy to model. Unfortunately, prosocial behavior is often more complex
and consequently less easily modeled.
The second factor influencing the modeling processes are the attributes of the
observers. People with low self-esteem, incompetent, or highly dependent are more
likely to engage in a modeled behavior (Engler, 1999). The third and strongest factor
determining the modeling processes are rewards associated with aggressive behavior.
When aggressive behavior of a model is rewarded the observer is more likely to engage
Violence and Aggression
13
in the same behavior as the model (Engler, 1999). The third factor explains why an
individual "with a history of successful goal attainment through aggressive means
develops both the expectation of reinforcement for aggressive behavior and the
necessary confidence to execute those behaviors" (Roach, 2002 p. 4) therefore
maintaining and further developing aggressiveness in society. An individual is more
likely to engage in aggressive behavior if he or she has been previously rewarded for
that behavior. For example a kid who is praised by his peers for winning a fight is more
likely to engage in the same behavior again because his peers previously rewarded
him. Furthermore, easily learned behaviors, such as violence and aggression are
learned more rapidly in an environment having high levels of violence and where
behaving aggressively is indirectly rewarded. The population's observational learning
process makes them more prone to engage in aggressive behaviors in order to
succeed.
Social Information Processing Model
The social information processing model of aggression is based on the idea
that the behavior of a person in a conflict is determined by the person's cognitive
represe,ntation of the event. The recent research proposed two main models of social
information processing, one by Huesmann in 1988 and the other by Crick and Dodge in
1994. Huesmann proposed a four step decision making model that children use to solve
social conflicts. The first step is to evaluate present environmental cues. The second
step is to search in one's memory for a script to guide behavior. The third step is to
evaluate the generated script, and finally the fourth step is to enact the script.
Huesmann's theory has its roots in earlier cognitive theories adopting the concept of
Violence and Aggression
14
scripts or guides to behavior generally stored in one's memory. The scripts are the
concepts that explain the maintenance of aggressive behaviors (Boxer & Dubow, 2001).
The second information processing model was proposed by Crick and Dodge
(1994). This model proposed that children have biologically limited capabilities and
scripts based on past experiences. Crick and Dodge's model consists of six steps. The
first step is the process of encoding and representing in memory the external and
internal cues related to the social context (Roach, 2002). During the encoding process
the individual's preexisting and reactive internal arousal affects the attention factors
which, along with the sum of previous experiences, determines their perception.
'
For
example if a person has been exposed to violence or child maltreatment they direct their
attention to hostile cues (Boxer & Dubow, 2002). The second step involves
interpretation of the encoded cues and making attributions about the intent of the actor.
The model argues for the possibility of a violent attribution bias in which the subject
attributes hostile motives to other's behaviors (Coccaro, 2003). The third step in Crick
and Dodge's model is the clarification and selection of the goals followed by the fourth
step, accessing the potential responses. The fifth step is to evaluate and choose a
response in relation to the desired outcome. The sixth and final step is to enact the
behavior.
The social information processing model suggests that aggressive responses
lower the possible access to future nonaggressive responses available in the
individual's reservoir of responses (Coccaro, 2003) thereby making an aggressive
individual more likely to engage in aggressive behaviors in the future.
Violence and Aggression
15
Religiosity and Aggression
There are a number of potential mediators of the relationship between
exposure to aggression/violence
and engagement in aggressive behavior. One such
mediator is the nature of one's religious faith. The underlying logic goes something like
this. Frustration can lead to aggression, as can attributions of aggressive intent. Thus,
adoption of belief systems that reduce frustration andlor decrease the likelihood of
making aggressive attributions of other's behaviors may decrease the likelihood of
engaging in aggressive behaviors.
Religiosity may play such a role.
Individuals can use religiosity as a coping mechanism to deal with stressful
situations, and crises, trauma, or life transitions (Pargament, Feuille, & Burdzy, 2011).
Some of the religious coping strategies include religious reappraisals, collaborative
relationships and spiritual support (Ano & Vasconcelles, 2005) that leads to spiritual
growth, positive affect and higher self-esteem.
Moreover, positive religious coping strategies portray a secure relationship
between the individual and a transcendent force, a spiritual connection with others and
a positive view of the world. On the other hand negative religious coping methods
suggests tension and struggle with the self, a transcendent force and other members of
the community (Pargament, Feuille, & Burdzy, 2011).
A study by Ano and Vasconcelles (2005) proposed four hypotheses for the
relationship between religious coping strategies and dealing with stressful situations.
The types of relations are positive religious coping with positive adjustment, positive
Violence and Aggression
16
religious coping with negative adjustment, negative religious coping with positive
adjustment and negative religious coping with negative adjustment.
Findings from this
study supported the hypotheses that positive religious coping is related to positive
psychological adjustment to stress (e.g., high self-esteem, life satisfaction and quality of
life) and that negative religious coping is related to negative psychological adjustment
(e.g., elevated depression and anxiety). Taken together, these findings suggest that
religious coping strategies are strongly associated with psychological adjustment to
stress (Ano & Vasconcelles, 2005).
Based on the previous research one potential mediator of the relationship
between exposure to aggression and engagement in aggressive behavior would be use
of positive and/or negative religious coping strategies.
Reactive and Proactive Aggression
Based on the aggression theories described above it is evident that aggression
is a multifaceted concept that can be roughly divided into two subtypes
reactive and
proactive. As described above, the frustration aggression theory established that
aggressive behavior is caused by a defensive reaction towards a perceived threat in
order to neutralize it. Therefore the defensive behavior is often evident in emotions like
anger and expressed in facial gestures and hostile verbalization.
This type of reactive
aggression is closely related to an idea of being out-of-control that escalates in intensity
and is often caused by a provoker. On the other hand, the proactive subtype of
aggression is consistent with the social learning theory in which aggression is
considered an instrument to attain a desired outcome and therefore controlled by
Violence and Aggression
17
reinforcements. Regularly proactive aggression is related to dominance and bullying
and usually occurs without an immediate provocation but is done to achieve a specific
goals (Roach, 2002).
Based on the two subtypes of aggression the current study seeks to investigate
the correlation between exposure to violence and the engagement in reactive or
proactive aggression considering the potential mediator of religiosity as a coping
mechanism. Considering the studies mentioned above higher levels of exposure to
violence were predicted to positively correlate with proactive aggression in college
students and either a positive or negative religious coping strategy could mediate the
strength of that relationship.
Method
Procedures
Participants in this study are currently enrolled Eastern Michigan University
students. The study was listed and described on the EMU electronic study management
system SONA. The participants were recruited through in-class announcements asking
for volunteers to participate in an anonymous study examining the relationship between
exposure to violence and feelings of aggression. Potential subjects initially accessed the
study via the SONA system and after choosing the study they were first taken to the
study's consent form, in which subjects clicked continue indicating they consent to
participate. After completing the informed consent form subjects were automatically
directed to a separate URL, survey monkey, in which the subjects will had access to the
Violence and Aggression
study questionnaires.
18
Completion of the study was estimated to take approximately 45
minutes but inspection of the time-to-completion
took 7.3 minutes to complete the study.
data indicate that subjects typically
The questionnaires administered in the study
include a brief demographic measure, followed by measures about exposure to
violence, reactive and proactive aggression, and religiosity. After subjects completed
the questionnaires they were linked to an external and independent URL in which they
could provide their name and the course for which they would like to receive extra
credit. The amount of extra credit received was determined by individual instructors.
Measures
Reactive and proactive aggression. The level of engagement in reactive and
proactive aggression was assessed by a six-item measure adapted from Chaux,
Arboleda and Rincon (2012). The first question assesses for propensity for retaliation
using a binary (yes/no) structure as follows "When you are treated badly, do you
retaliate?". The remaining items ask respondents to assess the frequency of
engagement in each aggressive behavior using the following scale: 0 = never, 1 =
almost never, 2 = almost always and 3 = always. To measure reactive aggression
participants were asked the following questions "When they treat you badly, do you
retaliate immediately" and "When they treat you badly, do you wait a while before
retaliating?". To measure proactive aggression participants were asked "Do you
threaten others to get what you want?", "Do you bully and make others feel bad?" and
"Do you enjoy treating others badly?". The measure yields separate subscale scores for
reactive aggression and proactive aggression.
Sub-scale scores were computed by
summing the individual item values within each scale. Thus, scores could range from 0
Violence and Aggression
19
to 6 for the two-item reactive aggression subscale and 0 to 9 for the three-item proactive
aggression subscale.
Higher scores on these scales reflect greater levels of proactive
and reactive aggression respectively.
Estimates of internal consistency reliability
(alpha) were computed for each subscale (reactive aggression: a = .03; proactive
aggression: a = .68.
Exposure to violence. The level of exposure to violence was assessed by an
adapted version of the Things I've Seen and Heard measure. The original version of the
scale included 15-items (Richters & Martinez 1992) assessing exposure to violence and
feelings of safety. The current paper used an adapted version of the Things I've Seen
and Heard measure that was developed as a part of the Longitudinal Studies of Child
Abuse and Neglect (Runyan et al. 2011). Items in the current version ask about
frequency of exposure to: community violence, violence-related activities (drugs,
arrests), violence within the horne, and direct experience of violence. The instrument
also asks about feelings of safety at home, in school, and with adults in general. The
measure yields two sub-scales assessing exposure to violence (16-items) and feelings
of safety (4 items). The violence exposure subsea Ie included questions like "I have seen
gangs in my neighborhood" and "I have seen somebody get shot". The items assessing
violence exposure ask respondents to estimate the frequency of exposure using the
following scale: 0
= never, 1 = once, 2 = twice, 3 = three times, 4 = four or more.
The
perceptions of safety subscale includes items such as the following: "I feel safe when I
am at home", "I feel safe when I am at school". The items on this subscale ask
respondents to estimate the frequency of perceptions of safety using the following
scale: 0
= never, 1 = almost
never, 2
= sometimes, 3 = almost
always, 4
= always.
Violence and Aggression
20
Sub-scales scores were computed by summing the individual item values within each
scale. Thus, scores could range from 0 to 64 for the 16-item violence exposure scale
and 0 to 16 for the four-item perceptions of safety scale. Higher scores on these scales
reflect greater levels of exposure to violence and greater perceptions of safety
respectively.
This measure is ideal for the current project because it assesses lifetime
exposure to community trauma, while excluding potentially distressing questions about
personal victimization. Estimates of internal consistency reliability (alpha) were
computed for each subscale (violence exposure: a = .81; feelings of safety: a = .73).
Religiosity. Religiosity was measured with the 14-item Brief RCOPE, which was
designed to evaluate how people use religion to cope with stress (Pargament, Feuille, &
Burdzy, 2011). The Brief RCOPE yields scores for two seven-item sub-scales
assessing positive religious coping and negative religious coping. Items for both subscales were scored on a 4-point Likert-type scale with the following anchors: 1 = not at
all, 2 = somewhat. 3 = quite a bit, and 4 = a great deal. The positive religious coping
subsea Ie included items such as the following: "Sought help from God in letting go of my
anger" and "Looked for a stronger connection with God" while the negative religious
coping subscale included questions like "Wondered whether God had abandoned me"
and "Questioned the power of God". Sub-scales scores were computed by summing the
individual item values within each scale. Thus, scores could range from 7 to 28 for each
subscale with higher scores reflecting greater use of positive and negative forms of
religious coping. Estimates of internal consistency reliability (alpha) were computed for
each subscale (positive religious coping: a = .96; negative religious coping: a = .87).
Violence and Aggression
21
Data analysis
The analysis was conducted using SPSS to assess direction and strength of
relationships among all predictors (violence exposure, perceptions of safety), mediators
(positive religious coping, negative religious coping), and outcomes (proactive
aggression, reactive aggression.
Tests of mediation were conducted using the process
macro in SPSS to determine if positive and negative religious coping mediates the
relationship between exposure to community violence, perceptions of safety and both
types of aggression. Mediation analyses include estimates of direct, indirect, and total
effects of predictors/mediators
on the outcomes.
Results
Participants
Demographic data for the sample are presented in Table 1 through Table 4.
Overall, the majority of the sample was female 76.4% and the predominant age group
ranged from 18 to 29 years old (see Table1). The majority of students were in either
their Freshman or Sophomore year of college (see Table 2). The most frequently
reported ethnic heritage in the sample was either white (66.7%) or African American
(13%) (see Table 3).The predominant religion reported was Christianity (35%) followed
by no religion (30.9%), Catholicism(20.3%)
and Islam (7.5%) (see Table 4).
Violence and Aggression
Table 1
Age of the sample
Age range
N
%
17 or younger
3
2.4
18-20
77
62.6
21-29
27
22.0
30-39
10
8.1
40-49
4
3.3
50- 59
1
0.8
60 or older
1
0.8
Total
123
100.0
22
Violence and Aggression
Table 2
Year in College
Class standing
N
%
Freshmen
42
34.1
Sophomore
33
26.8
Junior
19
15.4
Senior
23
18.7
5th year or beyond
6
4.9
Total
123
100.0
Table 3
Racial/ethnic Heritage of the sample
Racelethnicity
N
%
White
82
66.7
Black or African American
16
13.0
American Indian or Alaskan Native
1
0.8
Asian
6
4.9
Native Hawaiian or other Pacific
1
0.8
Islander
Hispanic
8
6.5
From multiple races
9
7.3
Total
123
100.0
23
Violence and Aggression
24
Table 4
Reported Religion
Religion
N
%
Protestantism
3
2.4
Catholicism
25
20.3
Christianity
43
35.0
Judaism
1
0.8
Islam
7
5.7
Buddhism
5
4.1
Hinduism
1
0.8
Inter/Non-denominational
1
0.8
Other
8
6.5
No religion (yes/no)
38/85
30.9/69.1
Note. Subjects could endorse more than one option.
Approximately half of the sample (51.2%) indicated that they retaliate when
treated badly.
However scores on the reactive aggression measure were quite low (M
= 2.4, SO = 1.1) (see Table 5). Similarly, the observed mean for proactive aggression
(M = 0.8 SO = 1.2) reflected really low levels of engagement in proactive aggression.
Level of violence exposure was in the low to moderate range (M = 17.1, (SO = 10.0).
Violence and Aggression
Conversely, perception of safety was quite high with ratings indicating that participants
almost always feel safe (M = 11.9 SO = 2.5) (see Table 5). Regarding the use of
religious coping, participants reported using positive religious coping somewhat
frequently (M
=
15.5 SO
= 6.7)
and infrequently using negative religious coping (M
10.6 SO = 4.2) (see Table 5).
Table 5
Included Measures
Measure
a
Reactive Aggression
.03
M
2.4
Proactive Aggression
.68
0.8
Exposure to Violence
.81
17.1
Perceptions of Safety
.73
11.9
Positive Religious
Coping
Negative Religious
Coping
.96
15.5
.87
10.6
SO
1.1
1.2
10.0
2.5
6.7
4.2
=
25
Violence and Aggression
26
Bivariate Relationships Among Predictors, Mediators, and outcomes
Correlational analyses were used to examine the relationships among violence
exposure, perceptions of safety, positive religious coping, negative religious coping,
proactive aggression and reactive aggression. The results indicate that feeling safe and
engagement in proactive aggression are negatively and significantly correlated r(121) =
-.20, P < 0.05. This suggests that participants who feel safe in their environment are less
likely to engage in proactive aggression. The second relevant correlation reveals that
positive and negative religious coping are positively correlated r(121) = .39, P < 0.001.
The result suggests that participants who engage in positive religious coping are also
likely to engage in negative religious coping. The remaining correlations were not
statistically significant.
Mediation Analysis
The direct and indirect effects of exposure to violence or perceptions of safety on
reactive or proactive aggression are estimated with a continuous X. The proposed
mediator, positive or negative religious coping, is regressed on exposure to violence or
perceptions of safety (X) to produce a, and reactive or proactive aggression is
regressed on both positive or negative religious coping and exposure to violence or
perceptions of safety, which yields band c', respectively (see Figure 1).
Violence and Aggression
Positive Religious
Coping
/
Negative Religious
Coping
Exposure to Violence
Perception of Safety
\
Reactive Aggression
"
Proactive Aggression:
Figure 1. Mediation Analysis
Eight tests of mediation were conducted alternating the six variables mentioned
above. In the service of comprehensiveness,
findings from all 8 tests of mediation are
presented in Tables 6 - 13). However, only two findings were significant, located in
models five and seven (see Tables 10 and 12).
Table 6
Model 1
Consequent
Antecedent
X(Violence
Y (Proactive
Aggression)
Coeff.
SE
M(Positive Religious
coping)
Coeff.
SE
p
A
.04
.06
>.05
p
c'
.00
.01
>.05
b
.00
.00
>.05
iz
1.03
.33
<.001
Exposure)
M(Positive
Religious
Coping)
Constant
14.86
1.20
<.001
R2 = .00
F(1, 121) = .36, P > .05
F(2, 120) = .26, P > .05
27
Violence and Aggression
Table 7
Model 2
Consequent
Antecedent
X(Violence
Y (Reactive
Aggression)
Coeff.
SE
M(Positive Religious
coping)
Coeff.
SE
p
A
.04
.06
>.05
p
c'
.01
.01
>.05
b
.00
.00
>.05
2.34
.30
<.001
Exposure)
M(Positive
Religious
Coping)
Constant
14.86
1.20
<.001
R2 = .00
F(2, 120) = .19, P > .05
F(1, 121) = 1.36, P > .05
Table 8
Model 3
Consequent
Antecedent
X(Violence
Y (Proactive
Aggression)
SE
Coeff.
M(Negative Religious
coping)
SE
p
Coeff.
A
.05
.04
>.05
p
c'
-.01
.01
>.05
b
.00
.01
>.05
i2
.63
.35
<.001
Exposure)
M(Negative
Religious
Coping)
Constant
i1
9.82
.75
<.001
R2 = .01
R2 = .01
F(1, 121) = 1.55, P > .05
F(2, 120) = .60, p > .05
28
Violence and Aggression
29
Table 9
Model 4
Consequent
Antecedent
X(Violence
Y (Reactive Aggression)
M(Negative Religious
coping)
Coeff.
SE
p
a
.05
.04
>.05
Coeff.
SE
p
c'
.01
.01
>.05
b
-.01
.00
>.05
2.30
.20
<.001
Exposure)
M(Negative
Religious
Coping)
Constant
9.82
.75
<.001
F(1, 121) = .35, P > .05
F(1, 121) = 1.55, P > .05
Table 10
Model 5
Consequent
Antecedent
X(Safe)
A
.05
.24
>.05
M(Positive
Religious
Coping)
Constant
Y (Proactive Aggression)
M(Positive Religious
coping)
Coeff.
SE
p
i,
14.86
2.91
<.001
F(1, 121) = .05, p > .05
Coeff.
SE
p
c'
-.10
.04
>.05
b
-.01
.01
>.05
2.09
.58
<.001
F(2, 120) = 2.62, p> .05
Violence and Aggression
Table 11
Model 6
Consequent
Antecedent
X(Safe)
Y (Reactive Aggression)
M(Positive Religious
coping)
SE
p
Coeff.
A
.05
.24
>.05
M(Positive
Religious
Coping)
Constant
11
14.86
2.91
<.001
Coeff.
SE
p
c'
.01
.04
>.05
b
-.01
.01
>.05
i2
2.23
.48
<.001
R2 = .00
R2 = .00
F(1, 121) = .05, P > .05
F(1, 121) = .13, P > .05
Table 12
Model 7
Consequent
Antecedent
X(Safe)
M(Negative
Religious
Coping)
Constant
Y (Proactive
Aggression)
Coeff.
SE
M(Negative Religious
coping)
SE
Coeff.
p
A
i1
-.19
12.91
.15
1.80
>.05
<.001
p
c'
-.09
.04
>.05
b
-.03
.00
>.05
i2
1.72
.63
<.001
R2 = .01
R2 = .04
F(1, 121) = 1.70, P > .05
F(2, 120) = 2.71, p> .05
30
Violence and Aggression
31
Table 13
Model 8
Consequent
Antecedent
X(Safe)
A
.15
-.19
>.05
M(Negative
Religious
Coping)
Constant
Y (Reactive
Aggression)
SE
Coeff.
M(Negative Religious
coping)
SE
p
Coeff.
h
12.91
1.80
R2
F(1, 121)
<.001
c'
.01
.04
>.05
b
-.01
.02
>.05
iz
2.30
.57
<.001
= .01
= 1.70, P > .05
p
= .00
120) = .09, P > .05
R2
F(2,
In model number five (see Table 10) multiplying a and b yields the indirect effect,
ab = 0.5(-.01) = -0.005. This indirect effect of -0.005 means that two participants who
differ by one unit in their reported perceptions of safety are estimated to differ by -0.005
units in their reported intentions to engage in proactive aggression as a result of the
tendency for those under relatively higher perceptions of safety to engage in more
positive religious coping (because a is positive), which in turn translates into lower
proactive aggression (because b is negative). This indirect effect is not statistically
different from zero, as revealed by a 95% BC bootstrap confidence interval that is not
above zero (-0.01 to 0.01) indicating that positive religious coping does not moderate
the relationship between perceptions of safety and proactive aggression.
Violence and Aggression
32
The direct effect of perceptions of safety, c' = -.10, is the estimated difference
in proactive aggression between two participants experiencing the same level of
positive religious coping but who differ by one unit in their reported perceptions of
safety. The coefficient is negative, meaning that the person with higher perceptions of
safety but who is equal in positive religious coping is estimated to be 0.10 units lower in
his or her reported proactive aggression.
The total effect of perceptions of safety on proactive aggression is derived
by summing the direct and indirect effects: c = c' + ab = -0.10 + (-0.005) = -0.105
Two participants who differ by one unit in perceptions of safety are estimated to differ by
-0.056 units in their reported proactive aggression. The negative sign means the person
under greater perceptions of safety reports lower proactive aggression.
The second significant finding is in model number seven (see Table 12) in which
the indirect effect ab= -0.19(-0.03)= 0.0057 means that the two participants who differ
by one unit in their reported perceptions of safety are estimated to differ by 0.0057 units
in their reported proactive aggression as a result of the tendency for those under
relatively more perceptions of safety to engage in less negative religious
coping(because a is negative), which in turn translates into lower proactive aggression
(because b is negative). However, these indirect effect is not statistically different from
zero, as revealed by a 95%
Be bootstrap
confidence interval that is not above zero (-
0.03 to 0.01) indicating that negative religious coping does not moderate the
relationship between perceptions of safety and proactive aggression.
The direct effect of perceptions of safety on proactive aggression is c' = -.09
which is consistent with model five and indicates that coefficient is negative, meaning
Violence and Aggression
33
that the person with higher perceptions of safety but who is equal in negative religious
coping is estimated to be 0.09 units lower in his or her reported proactive aggression.
The total effect of perceptions of safety on proactive aggression is derived by
summing the direct and indirect effects: c = c' + ab = -0.09 + (0.0057) = -0.0843.Two
participants who differ by one unit in perceptions of safety are estimated to differ by 0.0843 units in their reported proactive aggression. The negative sign means the person
under greater perceptions of safety reports lower proactive aggression.
Both, models five and seven, are consistent in the results that a person
under greater perceptions of safety reports lower proactive aggression and neither
positive or negative religious coping influence that relationship. The remaining six
analyses did not yield any significant results.
Discussion
The original hypothesis in the study stated that higher levels of exposure to
violence would predict of higher levels of aggression, and religiosity would mediate the
strength of that relationship. The findings from this study partially support the hypothesis
indicating that greater perceptions of safety (the mirror image of exposure to violence)
predicts lower proactive aggression.
However, neither positive nor negative religious
coping mediate that relationship. The results are partially consistent with previous
literature by Chaux et al. (2012) in which exposure to violence was significantly
correlated with both proactive and reactive aggression.
Lastly, the hypothesized mediation effect was not found in the current study. It
was hypothsized that either positive or negative religious coping would mediate the
Violence and Aggression
34
relationship between exposure to violence and aggression but the findings did not
indicate any effects. Previous literature by Greil et al(1989), Segall et. al (1989)and
Schuster et al.(2001) (as cited in Pargament, 2011) suggests that people engage in
religious practices as a strategy to deal with critical life situations. Moreover, Ano and
Vasconcelles (2005) found consistent links between measures of religious coping and
health and well-being indicators in different populations facing critical life events. As
exposure to violence and aggression are linked to crisis, trauma or life transitions it was
hypothesied that religiosity could have an effect on these variables as a strategy to cope
with challenging events. The mediation effect was not found in either postive or negative
religious coping.
There are three overarching limitations to the current study, which may have
limited the power to detect differences between groups. These limitations refer to the
limitations in characteristics of the sample, limitations in the measures used to assess
the core study constructs, and the failure to include alternative mediators. Turning first
to limitations in the sample, the current sample was drawn from a convenience sample
of college students, many of whom did not report elevations in violence exposure,
religious activities, or engagement in proactive or reactive aggression. One effect to
consider in the analyses is the limitation in the magnitude of the correlations due to a
restricted range of responses. The more limited the range of the responses the lower
the strength of the correlation and in the current study the range of the responses was
restricted by the characteristics of the sample. Perhaps one of the limitations of the
study leading to the non-significant findings is attributable to the questionnaires being
self-reports of aggression and violence in which participants may be particularly
Violence and Aggression
35
reluctant to endorse items related to engagement in more aggressive behaviors or
violent situations. This is consistent with the low mean scores on the aggression and
violence variables across analyses.
A potential option to increase the range of responses in the sample could be to
recruit participants with higher scores on aggression and who show at least some
religious preferences. These participants need to be recruited in settings other than
EMU's psychology classes. Recruiting the sample from higher crime rate cities and vary
religious community gatherings could yield more consistent results due the increase in
the response range.
Turning now to limitations in the measures used to assess core study constructs
the main limitation emerges regarding the reactive aggression questionnaire. The
measure was included in the current study because it was previously used by Chaux et
al. (2012) in a study assesing aggression and exposure to community violence. The
results for the measure in the current study indicated a very low internal consistency
reliability or alpha level of 0.03. The low level of reliability of this measure impacts all the
potential results related to this construct.
This limitation is consistent with lack of
findings related to reactive aggression in the study.
A third limitation relates to the absence of additional mediator variables that
might also impact the relationship between exposure to violence and aggression such
as emotion regulation or forgiveness. Emotion regulation was found to be a mediator
between interparental aggression and marital physical and emotional aggression
(Delsol, 2004). The previous mediation of emotion regulation could also have an effect
Violence and Aggression
36
on the relationship between level of aggression and exposure to community violence.
Another potential mediator is forgiveness.
Prior research has found that forgiveness
mediated the intergenerational transmission of violence (Rivera & Fincham, 2015).
Specifically, greater levels of forgiveness mediated the relationship between severity of
mother-perpetrated
violence and offspring dating violence victimization.
Future research should focus on further examining the effect of higher
perceptions of safety on lower levels of proactive aggression. Additional research could
study this relationship in samples characterized by higher exposure to violence and/or
more frequent engagement in religious practices. Lastly, as mentioned earlier, it is
important to evaluate other potential mediators of the relationship between perceptions
of safety and proactive aggression.
Violence and Aggression
37
References
Allwood, M. A, & Bell, D. J. (2008, November). A preliminary examination of emotional
and cognitive mediators in the relations between violence exposure and violent
behaviors in youth. Journal of Community Psychology,
Andreu, J. M., Arango-Lasprilla,
Porras, J., et al. (2003
989-1007.
J. C., Gonzalez, E., Castrill6n, D., Valencia, O. L.,
~lIIn13-March).
Diferencias culturales y de genero en la
representaci6n social de la agresi6n: Comparaci6n entre estudiantes espanoles
y colombianos. Avances en Psicologia Clinica Latinoamericana
, pp. 93-100.
Ano, G. G. (2005). Religious Coping and Psychological Adjustment to Stress: A MetaAnalysis. Journal of Clinical Psychology,
461-480.
Ardila, R. (2005). Aspects., Violence in Colombia: Social and Psychological. In R.
Ardila. Praeger Publishers/Greenwood
Publishing Group (Westport, CT, US).
Bell, P., & Baron, R. A (1979). Aggression and Heat: the mediating role of negative
affect. Journal of applied social psychology,
Berkowitz, L. (1989
~lIIn10-July).
18-30.
Frustration-aggression
hypothesis: Examination and
reformulation. Psychological Bulletin, pp. 59-73.
Boxer, P., & Dubow, E. F. (2001). A social-cognitive information-processing
model for
Violence and Aggression
38
school-based aggression reduction and prevention programs: Issues for research
and practice. Applied & Preventive Psychology,
177-192.
Bradshaw, C. P., Rodgers, C. R., Ghandour, L. A., & Garbarino, J. (2009). Socialcognitive mediators of the association between community violence exposure
and aggressive behavior. School Psychology Quarterly,
199-210 ..
Calvete, E., & Orue, I. (2011). The impact of violence exposure on aggressive behavior
through social information processing in adolescents. American Joumal of
Orthopsychiatry,
38-50.
Chaux, E., Arboleda, J., & Rincon, C. (2012). Community violence and reactive and
proactive aggression: The mediating role of cognitive and emotional variables ..
Revista Colombiana de Psicologia , 233-251.
Coccaro, E. F. (2003). Aggression:
psychiatric assessment and treatment. New York:
Marcel Dekker.
Crevecoeur, D. A. (2007
~IiIJl27-August).
Provocation, bias and arousal in hate crimes:
Cognitive neoassociation re-examined. Dissertation Abstracts International:
Section B: The Sciences and Engineering,
p. 668.
Delsol, C. (2004). Emotion regulation as a mediator between family-of-origin aggression
and marital aggression .. Pro Quest Information & Learning, 6325.
Engler, B. (1999). Personality theories: An introduction (5th ed.). Boston: Houghton,
Violence and Aggression
39
Mifflin and Company.
Hobfoll, S. E. (1988). The ecology of stress. Washington: Hemisphere Publishing Corp.
Johnson, K. A, Li, Y. J., Cohen, A B., & Okun, M. A (2013). Friends in high places:
The influence of authoritarian and benevolent god-concepts on social attitudes
and behaviors. Psychology of Religion and Spirituality,
15-22.
Leach, M. M., Berman, M. E., & Eubanks, L. (2008). Religious activities, religious
orientation, and aggressive behavior. Journal for the Scientific Study of Religion,
311-319.
Pargament, K., Feuille, M., & Burdzy, D. (2011). The Brief RCOPE: Current
Psychometric Status of a Short Measure of Religious Coping. Religions,
51-76.
Rivera, P. M., & Fincham, F. (2015). Forgiveness as a mediator of the intergenerational
transmission of violence. Journal of Interpersonal Violence, 895-910.
Roach, C. N. (2002 lI1l-1n 12-January). The distinctions of relational-overt and reactiveproactive forms of aggression: Social-adjustment correlates and predictors.
Dissertation Abstracts International: Section B: The Sciences and Engineering,
p.4922.
Scarpa, A, & Haden, S. C. (2006). Community Violence Victimization and Aggressive
Behavior: The Moderating Effects of Coping and Social Support. Aggressive
Behavior,
502-515.
Violence and Aggression
Schwartz, D., & Proctor, L. J. (2000). Community violence exposure and children's
social adjustment in the school peer group: The mediating roles of emotion
regulation and social cognition. Journal of Consulting and Clinical Psychology,
670-683.
Tate, W. (2011). Paramilitary Forces in Colombia. Latin American Research Review,
pp. 191-200,235.
Thomas, A., & Angelique, L. (2000). The role of peer support, individual attitudes, and
stress on aggression in urban African American adolescents. Dissertation
Abstracts International Section A: Humanities and Social Sciences, 491.
40